SBS, a France-based financial technology company serving more than 1,500 financial institutions, has launched SBS AI Foundation, a platform for running generative and agentic AI in core banking. Powered by NVIDIA, the platform allows banks to deploy AI without their data leaving their own environment, whether on a private cloud, a managed cloud, or fully on-premises. It is already in production with UK-based Scottish Building Society on Amazon Web Services and is now available to financial institutions across SBS’s client base.
Data residency and AI adoption in banking
The launch addresses a constraint many banks face when adopting AI. Regulations ranging from GDPR to country-specific data residency requirements limit how customer and transaction data can be processed, which prevents banks from routing sensitive data through external AI providers. Gartner data cited by SBS shows that while 84% of finance organisations have implemented or plan to implement AI, only 7% report a high or very high impact from those investments, a gap SBS attributes to the lack of infrastructure for running AI within a bank’s own systems.
The platform runs on NVIDIA infrastructure, with NIM microservices powering the AI layer, NeMo Guardrails keeping responses within each bank’s regulatory boundaries, and open Nemotron models. It draws data from across a bank’s SBS systems, including core banking, lending, digital banking, and regulatory reporting, and makes it available to generative AI and autonomous agents. According to SBS, the deployment at Scottish Building Society has delivered approximately 30% to 40% faster response times with more consistent outputs, while maintaining the explainability and audit trail required by regulators.
Three use cases and an agentic AI roadmap
SBS AI Foundation currently covers three areas. The first is AI-guided customer onboarding and engagement that uses a bank’s own records and product portfolio to identify suitable products for customers. The second is a natural language co-pilot that turns complex queries on structured and unstructured data into insights, replacing multi-day reporting cycles, while the third is risk and compliance monitoring that generates a full audit trail without sending risk data outside the institution’s controlled environment. SBS plans to add agentic AI capabilities, enabling autonomous agents to carry out multi-step core banking processes under human oversight.
Eric Bierry, CEO of SBS, said banks want AI they can trust, explain, and run wherever their data must be stored, and that building with NVIDIA allows SBS to provide this without compromising on governance. Aser Blanco, Global Head of Banking at NVIDIA, said SBS is turning a governed AI runtime into a shared intelligence layer across its banking platform, giving banks the tools to deploy generative and agentic AI across cloud and sovereign infrastructure.
